Data Mining

Data mining allows organizations to utilize information derived from a line of business and use that information to make predictions about future business trends. Data mining predictions can help a business make better decisions about its future direction and how to make the best use of its resources. SQL Server 2000 provided two basic data mining algorithms: decision trees and clustering. To these SQL Server 2005 adds several new data mining algorithms. The data mining algorithms that are included with SQL Server 2005 are: Decision Trees, Time Series, Sequence Clustering, Naïve Bayes, and Association Rules.

Decision Trees

The Microsoft Decision Trees (DT) algorithm is designed primarily for prediction. This algorithm is used to predict continuous and discrete variables.

Times Series

The new Time Series algorithm introduces the concept of past, present, and future into the prediction business. Designed to predict the next steps of the numerical sequence, the Time Series algorithm not only selects the best predictors for a given target but also identifies the best time periods over which you should expect to notice the effect of each predicting factor.

Clustering and Sequence Clustering

The Clustering algorithm is designed to find a good cluster count for your model given the properties of the training data. Sequence clustering allows you to find clusters of sequences of data. In other words, it is order-sensitive clustering.

Naïve Bayes

The new Naïve Bayes algorithm is a predictive algorithm. Designed for very fast performance, this algorithm predicts relationships between classifications of items.

Association Rules

Designed for analyzing transactional data, the Association Rules algorithm is used to find groups of items that exist within a single transaction.